****** Read Me Doc ******

This replication archive contains all documents (data, code) needed to replicate all results reported in the manuscript as well as in the Appendix. Questions regarding the replication material can be addressed to Maxime Walder (maxime.walder@unige.ch).

Please note that, since the paper employs Bayesian IRT models, the results obtained may vary slightly depending on the platform used (Windows, Mac, or Linux), but the conclusions are substantively identical. This likely has to do with the IRT model that relies on “rstan” which in turn relies on C++ (see here: https://mc-stan.org/docs/2_18/reference-manual/reproducibility-chapter.html). For reproducibility, you can rely on the models saved with the replication material

*** Structure

Four folders and one R-Project object. The folder ‘Code” contains all R scripts needed. The folder ‘Data” contains all datasets needed. The folder plots contains all the figures of the paper and the appendix, and the folder model contains all the objects with Bayesian IRT models saved in them. 

*** Materials included


** ‘Code’ Folder

- ‘Code/1. Run IRT models.R’ - Code that estimates the static and dynamic party competition spaces.
- ‘Code/2. Prepare municipal and cantonal data.R’ - Code that prepares the data for cantons and municipalities.
- ‘Code/3. Run models cantons and municipalities’ - Code that computes the position of cantons and municipalities in the party competition space.
- ‘Code/4. IRT models parties and population subsets’ - Code that estimates the party canton and party municipality space.
- ‘Code/5. Plots paper’ - Code that replicates the figures in the paper.
- ‘Code/6. Appendix plots’ - Code that replicates the figures in the appendix.

** ‘Plots’ Folder

- ‘Figure 1 -- Figure A16‘ - Figures in the paper and in the appendix.

** ‘model’ Folder

- ‘Simple IRT.stan‘ - Stan code for the static model.
- ‘Dynamic IRT.stan‘ - Stan code for the dynamic model.
- ‘Model party ideal position.rda‘ - Model results for the static IRT.
- ‘Model party ideal position_by legislature.rda‘ - Model results for the dynamic IRT.
- ‘Model IRT with parties and cantons.Rda‘ - Model results for the party mcanton IRT model.
- ‘Model IRT with parties and municipalities.Rda‘ - Model results for the party municipality IRT model.

** ‘Data’ Folder

-‘Data/Data model cantons and municipalities.Rda’ - List with dataset to estimate the static and dynamic position of cantons and municiplaities.
-‘Data/Results position municipalities and cantons.Rda” — List with random slop for cantons and municipalities, static and dynamic positions.
-‘Data/VZ_1888.xlsx’ - Dataset with census data from 1888 on population size, share of catholics, share of German speakers, and share of foreigners.